11.sla driven load balancing for web applications in cloud computing environment

Information and Knowledge Management www.iiste.org
ISSN 2224-5758 (Paper) ISSN 2224-896X (Online)
Vol 1, No.1, 2011

SLA Driven Load Balancing For Web Applications in Cloud
Computing Environment
More Amar
Computer Engineering Department, Pune University, MAE, Alandi
Pune, Maharashtra 412105, India
amarmore2006@gmail.com

Kulkarni Anurag
anurag.kulkarni@yahoo.com

Kolhe Rakesh
rakeshkolhe139@gmail.com

Kothari Rupesh
kotharirupesh@gmail.com

Yahide Prashant
yahideprashant@yahoo.com
Abstract
Cloud computing is an emerging topic in the field of parallel and distributed computing. Many IT giants such
as IBM, Sun, Amazon, Google, and Microsoft are promoting and offering various storage and compute
clouds. Clouds provide services such as high performance computing, storage, and application hosting.
Cloud providers are expected to ensure Quality of Service (QoS) through a Service Level Agreement (SLA)
between the provider and the consumer. In this research, we develop a heterogeneous test bed compute cloud
and investigate adaptive management of resources for Web applications to satisfy a SLA that enforces
specific response time requirements. We develop a system on top of EUCALYTPUS framework that actively
monitors the response time of the computed resources assign to a Web application and dynamically
allocates the resources required by the application to satisfy the specific response time requirements.

Keywords: Eucalyptus, SLA, QOS, Virtualization, Minimum Response Time.
1 Introduction

1.1 Overview:
Cloud Computing is an emerging topic in the field of parallel and distributed computing. Many IT giants
such as IBM, Sun, Amazon, Google, and Microsoft are promoting and offering various storage and
compute clouds. Clouds provide services such as high performance computing, storage, and application

10 | P a g e
www.iiste.org

Vol 1, No.1, 2011
hosting. Cloud providers are expected to ensure Quality of Service (QoS) through a service level
agreement (SLA) between the provider and the consumer. Cloud providers need to establish a robust
infrastructure that can grow dynamically, with easy maintenance and update.

A cloud is a combination of physically and virtually connected resources. Virtualization allows us to
instantiate virtual machines dynamically on physical machines and allocate them resources as needed.
Therefore, virtualization is one of the key technologies behind the cloud computing infrastructure. There
are several benefits that we expect from virtualization, such as high availability, ease of deployment,
migration, maintenance, and low power consumption that help us to establish a robust infrastructure for
cloud computing.

EUCALYPUTS is an open source framework for developing clouds. Its primary developers are at the
University of California, Santa Barbara. The aim of EUCALYPTUS is to enable academics to perform
research in the field of cloud computing. In this research study, I use the EUCALYPTUS framework to
establish a cloud and host a simple Web application on a Web farm of virtual machine instances.

1.2 Problem Statement

In Cloud computing, provider and consumer sign a service level agreement (SLA) that defines quality of
service requirements. Cloud providers such as Amazon and Google allow consumers to the select resources
they need for their cloud-based applications. Usually consumers reserve resources without any optimal
prediction or estimation, leading either under or overutilization of reserved resources. Web application
owners, in order to ensure the us- ability of their applications, need to maintain a minimum response
time for their end users. Web application owners should be able to host their applications on a cloud by
specifying a desire Quality of Service (QoS) in terms of response time. Cloud providers should ensure
those QoS requirements using a minimal amount of computational resources. When violations of the
SLA are predicted or detected, cloud providers should horizontally scale up the hosted Web application
to ensure that the specific response time requirement is satisfied. Currently, neither the commercial cloud
providers nor the existing open source frameworks support maximum response time guarantees.

1.3 Objectives

The main goal of this research study is to satisfy a cloud consumer that specifies Quality of Service (QoS)
requirements in terms of response time by monitoring the average response time of the application and
adaptively scaling up the Web application when necessary. Towards reaching this goal, the following are
o u r specific objectives:
1. Understand EUCALYPTUS and establish a heterogeneous compute cloud test bed.
2. Write a simple Web based application and develop a workload generator for it.
3. Host the Web application on the cloud and establish a load balancer for the application.
4. Detect SLA violations by monitoring request response times.
5. Develop software components to horizontally scale up the Web application dynamically.

2 Literature and Technology Review

2.1 Cloud Computing
Cloud computing is an emerging technology in the field of parallel and distributed computing. Cloud
computing is the delivery of computing as a service rather than a product, whereby shared resources, software,
and information are provided to computers and other devices as a utility over the Internet.
Definition:“A Cloud is a type of parallel and distributed system consisting of a collection of interconnected
and virtualized computers that are dynamically provisioned and presented as one or more unified

11 | P a g e
www.iiste.org

Vol 1, No.1, 2011
computing resources based on service-level agreements established through negotiation between the service
provider and consumers”.
Storage and compute clouds are two major types of clouds that aim to provide services without exposing
the underlying infrastructure to customers. A cloud service has three distinct characteristics that differentiate
it from traditional hosting. It is sold on demand, typically by the minute or the hour; it is elastic -- a user can
have as much or as little of a service as they want at any given time; and the service is fully managed by the
provider (the consumer needs nothing but a personal computer and Internet access). Significant innovations in
virtualization and distributed computing, as well as improved access to high-speed Internet and a weak
economy, have accelerated interest in cloud computing. Instead of investing and managing personal hardware,
users rent virtual machines or services from cloud providers.

2.2 Open source Cloud framework (Eucalyptus)

2.2.1 EUCALYPTUS:
EUCALYPTUS is an open source cloud computing framework developed by the University of California,
Santa Barbara as an alternative to Amazon EC2. The aim of EUCALYPTUS is to enable academics to
perform research in the field of cloud computing. It is composed of several components that interact with
each other with well-defined interfaces. EUCALYPTUS therefore allows the open source community to
replace or modify any component according to their requirements. Key functionalities of EUCALYPTUS
are VM instance scheduling, construction of virtual infrastructure, administrative interfaces, user
management of clouds, and definition and execution of service-level agreements. EUCALYPTUS allows
the use of Amazon EC2 interfaces for interacting with the cloud. The three main components of
EUCALYPTUS are the Node Controller, the Cluster Controller and the Cloud Controller. Each is described
below.

2.2.1.1 Node Controller:
The Node Controller (NC) is responsible for controlling the execution, inspection, cleanup, and
termination of virtual machine instances on a physical machine. A EUCALYPUTS cloud may
contain many NCs, but one physical machine only requires one NC. A single NC is therefore able to
control all virtual machines executing on a single physical machine.

2.2.1.2 Cluster Controller:
A collection of NCs form Cluster Controller (CC) that typically executes on a head node or server that is
able to access both the public and the private network. The CC is responsible for managing a collection
of associated NCs.

2.2.1.3 Cloud Controller:
Each EUCALYPTUS-based cloud must include one Cloud Controller (CLC) that is the entry point for
users. The CLC is responsible for making global decisions that include user authentication, processing of
user or administrative requests, high level decisions for VM instance scheduling, and processing of
SLAs

2.3 Virtualization:
Virtualization allows multiple operating systems to execution simultaneously on a physical machine. It is
achieved through a virtual machine monitor (VMM). VMMs are also known as hypervisors. VMMs are
responsible for keeping track of all activities performed by virtual machines. VMMs are categorized into
Type-1 and Type-2 VMMs. Type-1 VMMs run directly on hardware without the need for a host OS,
while Type-2 VMMs run on top of a host OS. Server virtualization can be achieved using Type-1 or
Type-2 VMMs. Virtualization achieved through a Type-1 VMM is known as paravirtualization, and
virtualization achieved through a Type-2 VMM is known as full virtualization.
On the X86 architecture, four privilege levels are available and known as Ring 0, 1, 2, and 3. Privilege
levels are used to manage hardware resources to the applications and OS. Ring 0 is the most privilege

12 | P a g e
www.iiste.org

Vol 1, No.1, 2011
level and Ring 3 is the least privilege level. Usually, OS instructions execute in Ring 0 and user
application instructions execute in Ring 3. Virtualization needs to place VMM in the most privilege level
(Ring 0) for managing the share resources for multiple operating systems.

2.3.1 Virtualization platform KVM:
The Kernel-based Virtual Machine (KVM) is a free virtualization platform licensed under GPL version 2,
provides full virtualization. KVM contains a Type-2 hypervisor and supports Linux-based host operating
systems. It allows Windows or Linux as the guest OS. The Ubuntu community provides KVM support.
A typical KVM installation consists of the following components:
• A device driver for managing the virtualization hardware; this driver exposes its capabilities via a
character device /dev/kvm.
• A user-space component for emulating PC hardware; currently, this is handled in the user space and is
a lightly modified QEMU process.
• The I/O model is directly derived from QEMU's, with support for copy-on-write disk images and other
QEMU features.

3 System working and methodology

3.1 System Overview:
We develop a private network using four workstations to establish a heterogeneous cloud on the CS&IM
LAN. We used the EUCALYPTUS framework to establish a cloud with one Cloud Controller (CLC), one
Cluster Controller (CC), and three Node Controllers (NCs). We installed the CLC and CC on a front-end
node that is on the CSIM LAN and the clouds private network. We installed the NCs on three separate
machines connected to the private network. We used EUCALYPTUS to instantiate virtual machines on the
node controllers.
Ngnix is a HTTP and mail proxy server capable of load balancing .We used Ngnix as a load balancer
because it offers detailed logging and allows reloading the configuration file without termination of
existing client sessions. We installed the Ngnix as a load balancer on the front- end node and the Tomcat
application server on the virtual machine. W e stored the virtual machine image in EUCALYPTUS,
which cached it on all three NCs. Our Web application contains one Java servlet that performs a matrix
calculation. The Tomcat server hosts the Web application while Ngnix balances the load. We used Httperf
to generate a HTTP workload for the Web application. We developed a software component named
VLBManager to monitor the average response time of requests by real-time monitoring of Ngnix logs. We
developed another software component named VLBCoordinator to invoke the virtual machines on the
EUCALYPTUS-based cloud. Both components interact with each other using XML-RPC. The
VLBManager detects the violation of the average response time for each virtual machine in the Web farm
due to the heterogeneous testbed cloud. Whenever VLBManager detects a violation of the average
response time, it signals the VLBCoordinator to invoke another virtual machine and add it to the Web
farm.

3.2 System D esign:

3.3 Experiment: Adaptive resource allocation to Web application:
In this experiment, w e used our proposed system to prevent response time increases and rejection of
requests by the web server. . Initially, the load balancer is deployed with one virtual machine (vm1), while
vm2 and vm3 are cached to the physical system using eucalyptus. In this experiment, we try to satisfy a
service level agreement that enforces a two-second maximum average response time requirement for the
sample web application on an undefined load level. W e use vlbmanager to monitor the ngnix logs and

13 | P a g e
www.iiste.org

Vol 1, No.1, 2011
detect the violations of the sla. W e use vlbcoordinator to adaptively invoke additional virtual machines
as required to satisfy the sla.

3.4 Results:
The figure 3.3.1 shows the number of requests served by the system. We observe linear growth in the
number of requests served by the system with each load level. Figure3.3.2 shows the average response
time of each virtual machine as it is adaptively added to the load balancer. Whenever the system detects a
violation of the response time requirement from any virtual machine, it dynamically invokes another virtual
machine and adds it to the Web farm. We observe continued violation of the required response time due to
the latency of virtual machine boot-up.

3.5 Discussion:
Adaptive management of resources on compute clouds for Web application helps us offer
SLAs that enforce specific response time requirements. To avoid the continued violation of the SLA during
VM boot-up, it would be better to predict violation of response time rather than to wait until the
requirement is violated.

4 Conclusion and Recommendations

4.1 Overview:
Cloud computing is an emerging topic in the field of parallel and distributed computing and requires a
huge contribution from the research community. This research study explored cloud computing and
identified the advantage of adaptive resource management in heterogeneous clouds for Web applications.
We developed a system on top of a EUCALYPTUS cloud that adaptively grows to satisfy specific
response time requirements for Web applications.

4.2 Contribution:
This research identified a new kind of Service Level Agreement (SLA) for heterogeneous compute clouds in
which the consumer wants specific response time for his or her Web application. We developed a system on
top of EUCALYPTUS that actively monitors the response time of a Web application hosted on multiple
virtual machines and adaptively manage the Web application’s resources.

4.3 Recommendations
The CPU is typically the bottleneck in dynamic content generation and We observed that whenever
response time starts to grow, the CPU usage exceeds above 90%. We recommend investigating system
behavior by adaptively invoking other virtual machines by actively monitoring this behavior. To help
overcome the virtual machine boot-up latency problem it would be useful to predict VM response times in
advance then invoke additional virtual machines before the system exceeds the response time limit. It
should not be necessary to check each VM response time if load balancing is fair instead of round robin.
We would only need to check aggregated response time.

We also recommend moving VLBManager and Ngnix to a virtual machine and naming this VM as
the head node. Whenever a user needs to deploy a Web application on cloud, the provider needs to
deploy two virtual machines, one for the Web application and one for the head node to actively monitor
the response time of the Web application hosted on the virtual machine. The head node can detect
violations of response time requirements and adaptively deploy more virtual machines to satisfy the
response time requirements.

References:

14 | P a g e
www.iiste.org

Vol 1, No.1, 2011
Postfix server configuration
http://www.server-world.info/en/note?os=Fedora_11&p=mail
Creation of certificate
http://www.g-loaded.eu/2005/11/10/be-your-own-ca/
Cloud Computing
http://en.wikipedia.org/wiki/Cloud_computing
http://searchcloudcomputing.techtarget.com/definition/cloud-computing
http://www.ibm.com/cloud-computing/us/en/
Eucalyptus and KVM configuration
http://open.eucalyptus.com/wiki/EucalyptusInstallationSource_v2.0
http://www.howtoforge.com/virtualization-with-kvm-on-a-fedora-11-server

Avi, K., Yaniv, K., Dor, L., Uri, L., & Anthony, L. (2007). KVM: The Linux Virtual Machine Monitor
in Ottawa Linux Symposium (pp. 225–230).
Buyya, R., Yeo, C. S., & Venugopal, S. (2008). Market-oriented cloud computing: Vision, hype, and
reality for delivering it services as computing utilities. In HPCC ’08: Proceedings of the
2008- 10th IEEE International Conference on High Performance Computing and c ommunications
(pp. 5–13). Washington, DC, USA: IEEE Computer Society.
Google Code. (2008). Typica: A Java client library for a variety of Amazon Web services. Available at
http://http://workspace.globus.org// [Online; accessed 06-March-2009])
Wikipedia. (2009b). Google App Engine. (Available at http://en.wikipedia.org/w/index.
php?title=Google_App_Engine&oldid=273555538 [Online; accessed 3-March-2009])
VMWare2007). Understanding full virtualization, paravirtualization, and hardware assist.Whitepaper.
Amazon.com, Inc. (2009). Amazon Elastic Compute Cloud (EC2). (Available at
http://aws.amazon.com/ec2/ [Online; accessed 1-March-2009])

Fig2.2.1: Hierarchical interaction of Eucalyptus components

15 | P a g e
www.iiste.org

Vol 1, No.1, 2011

Fig3.2: Final System Architecture

Table: Hardware configuration of physical machines

Node Name Type CPU Frequency RAM

Front-end Intel Pentium 2.80 GHz 2 GB

Node1 Intel Pentium 2.66 GHz 1.5 GB

Node2 Intel Celeron 2.4 GHz 2 GB

Node3 Intel Core 2 Duo 1.6 GHz 1 GB

16 | P a g e
www.iiste.org

Vol 1, No.1, 2011

Figure3.3.1: Number of requests served by system during Experiment. We observe linear growth in the
number of requests served by system with each load level.

Figure3.3.2: Average response time for each virtual machine during Experiment. Whenever the
system detects a violation of the response time requirement from any virtual node, it dynamically invokes
another virtual machine and adds it to the Web farm.

17 | P a g e
www.iiste.org

International Journals Call for Paper
The IISTE, a U.S. publisher, is currently hosting the academic journals listed below. The peer review process of the following journals
usually takes LESS THAN 14 business days and IISTE usually publishes a qualified article within 30 days. Authors should
send their full paper to the following email address. More information can be found in the IISTE website : www.iiste.org

Business, Economics, Finance and Management PAPER SUBMISSION EMAIL
European Journal of Business and Management EJBM@iiste.org
Research Journal of Finance and Accounting RJFA@iiste.org
Journal of Economics and Sustainable Development JESD@iiste.org
Information and Knowledge Management IKM@iiste.org
Developing Country Studies DCS@iiste.org
Industrial Engineering Letters IEL@iiste.org

Physical Sciences, Mathematics and Chemistry PAPER SUBMISSION EMAIL
Journal of Natural Sciences Research JNSR@iiste.org
Chemistry and Materials Research CMR@iiste.org
Mathematical Theory and Modeling MTM@iiste.org
Advances in Physics Theories and Applications APTA@iiste.org
Chemical and Process Engineering Research CPER@iiste.org

Engineering, Technology and Systems PAPER SUBMISSION EMAIL
Computer Engineering and Intelligent Systems CEIS@iiste.org
Innovative Systems Design and Engineering ISDE@iiste.org
Journal of Energy Technologies and Policy JETP@iiste.org
Information and Knowledge Management IKM@iiste.org
Control Theory and Informatics CTI@iiste.org
Journal of Information Engineering and Applications JIEA@iiste.org
Industrial Engineering Letters IEL@iiste.org
Network and Complex Systems NCS@iiste.org

Environment, Civil, Materials Sciences PAPER SUBMISSION EMAIL
Journal of Environment and Earth Science JEES@iiste.org
Civil and Environmental Research CER@iiste.org
Civil and Environmental Research CER@iiste.org

Life Science, Food and Medical Sciences PAPER SUBMISSION EMAIL
Journal of Biology, Agriculture and Healthcare JBAH@iiste.org
Food Science and Quality Management FSQM@iiste.org
Chemistry and Materials Research CMR@iiste.org

Education, and other Social Sciences PAPER SUBMISSION EMAIL
Journal of Education and Practice JEP@iiste.org
Journal of Law, Policy and Globalization JLPG@iiste.org Global knowledge sharing:
New Media and Mass Communication NMMC@iiste.org EBSCO, Index Copernicus, Ulrich's
Journal of Energy Technologies and Policy JETP@iiste.org Periodicals Directory, JournalTOCS, PKP
Historical Research Letter HRL@iiste.org Open Archives Harvester, Bielefeld
Academic Search Engine, Elektronische
Public Policy and Administration Research PPAR@iiste.org Zeitschriftenbibliothek EZB, Open J-Gate,
International Affairs and Global Strategy IAGS@iiste.org OCLC WorldCat, Universe Digtial Library ,
Research on Humanities and Social Sciences RHSS@iiste.org NewJour, Google Scholar.

Developing Country Studies DCS@iiste.org IISTE is member of CrossRef. All journals
Arts and Design Studies ADS@iiste.org have high IC Impact Factor Values (ICV).

11.sla driven load balancing for web applications in cloud computing environment

Recommandé

Recommandé

Contenu connexe

Tendances

Tendances (14)

En vedette

En vedette (8)

Similaire à 11.sla driven load balancing for web applications in cloud computing environment

Similaire à 11.sla driven load balancing for web applications in cloud computing environment (20)

Plus de Alexander Decker

Plus de Alexander Decker (20)

Dernier

Dernier (20)

11.sla driven load balancing for web applications in cloud computing environment